Solution Icon

Solutions

Cutting-edge solutions tailored to everyday issues in the modern digital environment.

Services icon

Services

Innovative tools and services designed to address complex challenges today’s digital landscape.

IT FOR What WE DO MENU

Aerial view of an industrial facility at sunset

Services

Hexagon is reshaping industries through smarter systems, streamlined workflows, and bold digital innovation.

IT FOR HOW WE DO MENU

Aerial view of an industrial facility at sunset

Services

Hexagon is reshaping industries through smarter systems, streamlined workflows, and bold digital innovation.

Ready to clean up your CRM workflow?

Book a CRM Workflow Audit and see where leads, follow-ups, reporting, or disconnected tools can be fixed first.

Cloud Security

Azure Access Reviews: Prove Who Can Change Your Systems

Check every role against its scope, give app identities their own review, and remove access in tested steps.

  • Azure
  • Access Management
  • Updated October 2026
Dark cover with the words Azure Access Who Can Change What in white and blue

An Azure access review proves which people and applications can change which resources, and why they still need that access. We start with one business system, check every role against its scope, review application identities separately, and remove access in small, tested steps so nothing important breaks.

A list of account names does not answer those questions. This guide walks through the review we recommend for companies that run line-of-business systems on Azure. It is a practical workflow, not a promise that one checklist makes an environment secure or compliant.

Where should an Azure access review start?

Start with one production workload, not the whole subscription. Name its business owner, its technical owner, its data stores, its deployment process and the services it depends on. Then list every person, group, service principal and managed identity involved in running it.

Separate everyday use from administration. Someone who approves invoices in an application rarely needs to change the cloud resources behind it. A deployment pipeline may need to update the app without being able to manage every service in the subscription.

For each important grant, record five things: the action, the scope, the owner, the reason and the next review date. Anything without a clear reason gets flagged for investigation, not deleted on the spot. An unfamiliar identity may be holding up a scheduled job that nobody remembers.

How do we judge whether a role is too broad?

Look at the role and the scope together. Microsoft's Azure RBAC best practices recommend granting only the access people need, keeping a maximum of 3 subscription owners, scoping privileged roles to a resource group or resource where possible, and assigning roles to groups instead of individual users.

We turn that guidance into review questions rather than a bulk-removal script:

  • Does this support engineer need subscription-wide rights, or a narrower role on one resource group?
  • Does this admin need permanent access, or would time-limited access approved through Microsoft Entra Privileged Identity Management cover it?
  • Do any custom roles use wildcards that could quietly grant new operations later?
  • Who approves an exception, and when is it reviewed again?

Check inherited access, not just direct assignments. A resource can look locked down while a broader scope above it grants far more. Also keep resource management rights separate from data access rights inside the service, and test the exact operation each person or app needs.

Why do application identities need their own review?

Because apps hold some of the most powerful access in many environments, and nobody logs in as them to notice. Microsoft's managed identity recommendations say an identity that only reads from a storage account should not have write access, and warn against broad roles such as subscription-level Contributor unless they are truly required.

We keep a separate inventory for these identities. Each row names the workload, the resource it touches and the exact operation, such as reading one configuration value or writing results to one container. The owner is a named team or person who can explain the dependency, not just the IT department. If several apps share one user-assigned identity, we document that before changing anything, because one restriction will affect all of them.

Two details catch teams out. First, deleting a managed identity does not delete its role assignments; the orphans show as "Identity not found" in the portal and should be cleaned up. Second, managed identity tokens are cached for around 24 hours, so a permission change can take several hours to take effect. Plan test windows with that in mind.

How do we remove access without breaking the business?

Every proposed reduction gets a short change record: current state, intended state, expected impact, test and rollback. The workload owner signs off before it is applied.

Test in a representative non-production environment where you can. Confirm both sides: the allowed action still works and the blocked action now fails. A reporting service, for example, should still read its dataset and should fail if it tries to write. Use approved test accounts and data, never live customer records.

Roll changes out in small groups and watch normal usage, background jobs and deployments. A job that runs weekly will not be proven by a five-minute check, so the observation window has to cover its real cycle. If something fails, do not hand out a broad admin role to make the error go away. Find the exact denied operation, decide whether it is really needed, and write the answer down for the next reviewer.

What about emergency and vendor access?

Keep a documented emergency path owned by your security and operations leads. It needs named authorized users, a way to detect when it is used and a follow-up review afterward. Do not remove or redesign emergency access as part of routine cleanup without an approved plan.

Record the access needed for incidents, vendor support and deployment recovery too. These are real operating needs. The goal is to make them visible and proportionate instead of leaving permanent exceptions that nobody can explain. The same thinking applies to AI tools; we cover it in task-scoped AI agent permissions.

How do we measure a good access review?

Not by the number of assignments removed. A review can delete hundreds of grants and still leave the riskiest access in place. We track:

  • The share of important grants with a confirmed owner and reason.
  • Broad privileges that are still unresolved.
  • Exceptions past their review date.
  • Business tests completed after each change.

We also log any disruption the review itself caused. That feedback improves the next round. The end state is an access model your team can keep current as staff, vendors and applications change. Review on a schedule that fits each workload's risk, and also when people leave, roles change or an application is retired. If you are moving content into Microsoft 365 at the same time, our guide to SharePoint document management covers ownership on that side.

Frequently asked questions

What is an Azure access review?

It is a structured check of who and what can act on your Azure resources, at which scope, and why. It covers people, groups, service principals and managed identities, and ends with a tested plan to remove access that is not needed.

Is a managed identity automatically least privilege?

No. A managed identity removes stored secrets, but its access is only as narrow as the roles you assign to it. Review its roles like any other identity.

Should we remove every unexplained assignment right away?

No. Find the owner and dependency first, then remove it through an approved, tested change. An active security incident is different and should follow your security team's response plan.

How often should we review Azure access?

Set a regular cadence based on each workload's risk and rate of change, and also review when people leave, responsibilities change or an application is retired.

Sources

  1. Microsoft Learn: Best practices for Azure RBAC
  2. Microsoft Learn: Managed identity best practice recommendations

Map the access behind one business system

Bring us a workload diagram and a sanitized access list (no passwords or keys). We will help you sort what each identity really needs and build the change plan. Our custom development and API integration teams can then tighten the apps and integrations themselves.

BUSINESS SYSTEMS

SOFTWARE

AI

INTEGRATIONS

Build, Improve, or Connect the Systems Your Business Depends On.

Whether you need custom software, CRM, ERP, AI, automation, integrations, or modernization, tell us what you’re trying to improve. We’ll review your requirements and recommend the most practical next step.

300+

Software projects delivered

37+

Enterprise apps built

150+

delivery team members

★★★★★ 4.7 on Google · 20 reviewsU.S.-led since 200721 awards & certifications →
Client results
BMG: cut the time lost on follow-ups by 70% with our CRM.
Utah Dental Association: convention registration streamlined from months to one month.

Tell Us About Your Project

No sales pressure. Tell us what you’re trying to build, improve, integrate, automate, or replace. We’ll review your answers and come back with a clear recommendation within 1 business day.

Prefer to talk now? Call +1 (801) 259-4800 or book a free 30-minute review.

© 2026 Hexagon IT Solutions. All rights reserved.